I can see some potential value add in this space. Whenever you're going somewhere new and/or you don't already know your exact itinerary upfront, you would (nowadays) typically ask an AI bot for recommendations, and those are typically poorly integrated, and often AI will suggest places that don't exist, are closed etc., then you have to manually book them. I haven't checked travel apps space lately, but I believe there's room here to deliver a more seamless experience. Getting users is going to be the hard part.
Getting users is going to be the hard part.
It is certainly a hard part, but delivering value to those users will be even harder.
Because delivering value will require an LLM with expertise on every place a person might want to travel and delivering useful answers to people with arbitrary tastes/desires/needs/itineraries/etc.
Some times for some people in some places "McDonalds" is just the right recommendation about where to eat. If you are at Exit 47 at 9pm it could be the best option. If you are in the French Quarter at 3pm, it probably isn't.
So how do you pay for good answers for everyplace in the world? Even Google struggles with that.